Hierarchical Radial Basis Function Neural Networks for Classification Problems

Hierarchical neural networks consist of multiple neural networks assembled in the form of an acyclic graph. The purpose of this study is to identify the hierarchical radial basis function neural networks and select important input features for each sub-RBF neural network automatically. Based on the pre-defined instruction/operator sets, a hierarchical RBF neural network can be created and evolved by using treestructure based evolutionary algorithm. This framework allows input variables selection, over-layer connections for the various nodes involved. The HRBF structure is developed using an evolutionary algorithm and the parameters are optimized by particle swarm optimization algorithm. Empirical results on benchmark classification problems indicate that the proposed method is efficient.